Biological Conceptions of Race 459not share with the members of any other race — and this collection of traits is
supposed to be passed on faithfully in reproduction. Because many of the traits
typically used to individuate races (skin color, bone structure, etc.) are multigenic
and are most likely spread all over the human genome, independent assortment is
thought to call into question the assumption that races breed true to type.
The central problem with this argument rests in the assumption that the genes
for racial traits must be genetically linked in order for races to breed true to type.
Races will breed true to type even when racial traits are independently assorted,
provided that there is little outbreeding among racial groups. Inbreeding reduces
heterozygosity in the gene pool, thus making it more likely that traits will be passed
on faithfully and as a package in reproduction. Cultural factors and geographic
separation are two factors that might contribute to reproductive isolation.
Other critics of racial typology argue that races cannot be types because the
morphological traits typically used to individuate races (e.g., skin color, hair type,
eye shape, etc.) are often clinal and discordant [Banton and Harwood, 1975; Ap-
piah, 1996; Zack, 1997].^9 Not only do many so-called racial traits vary gradually
across a geographic region; they also vary independently. For example, a classifi-
cation scheme based solely on skin color might fail to agree reasonably with one
based solely on hair type — and both might fail to agree reasonably with one based
solely on bone structure, and so forth. Critics of racial typology sometimes add
that a similar problem can be found at the genetic level. Support for the latter
claim comes from Lewontin [1972] who argues that there is more genetic variation
within than among major racial groups (blacks, whites, and Asians). The conclu-
sion that is frequently drawn is that racial typology is false because these data cast
doubt upon the typological assumption that the members of a race have a cluster
of properties in common that they do not share with members of other races.
While this argument against racial typology is an improvement over the previous
one, it is still not fatal. As Sober has argued [1980], variation within and continuity
among taxa need not pose a problem for typological definitions in systematic
biology. By appeal to what Sober calls the ‘natural state model’, pre-Darwinian
essentialists had a way of explaining nature’s diversity that is consistent with
essentialism. According to this model, the members of a kind have a common
natural state, determined by the kind-specific essence, but interfering forces often
prevent individuals from realizing that state. In other words, essential properties
can be thought of as dispositional properties that, when there are no interfering
forces at work, have the propensity to manifest the traits typical of the members of
a kind. By appeal to this model, then, a racial typologist can recognize significant
variation within, and continuity among, races provided that there are discrete
natural tendencies underlying this variation.
Not only does this reply reveal a problem with the previous argument, it also
points to a better argument against racial typology. It is not variationper sethat
(^9) This argument combines some elements of the previous two arguments against racial typology
but also differs from them in that it makes no assumption about the cause of clinal and discordant
variation.